New Zealand Law Students Journal
Last Updated: 20 July 2016
REBUILDING BABEL: NEGOTIATING MEANING IN MULTILINGUAL LEGISLATION
SARAH J REESE
All law is, to an extent, a process of translation. Even when operating within the same language, the enactment and enforcement of any given piece of legislation requires first that the legislature, after determining what it means to enact, gives the relevant instructions to the drafters for them to ‘translate’ or ‘transform’ into written text.1 That legislation must then be promulgated, and applied to the real world by the judiciary through interpretation. In both cases, what occurs is in the broadest sense a translation, in that the communication of ideas between individuals can never be performed with complete precision: there is always some small gap between what is said and what is meant, and what is meant and what is understood, and the aim of good legislative drafting can only ever be to reduce these gaps as much as possible.2 In a way, then, many of the problems of law and legal interpretation find their echo in those which have preoccupied translators since the Classical period—the tension between the literal or “word-for-word” adherence to authority, and the more figurative (or purposive) “sense-for-sense” translation, which in legal terms implies the spirit rather than the letter of the law.3 This inherent conflict is
1 Robert Stanton, The Culture of Translation in Anglo-Saxon England (D S Brewer, Cambridge, 2002) at 4: “all translation is transformation.”
2 For further discussion on the idea that communication involves gaps, see John Berger and Jean Mohr, Another Way of Telling (Vintage International, New York, 1995).
3 Susan Bassnett Translation Studies (Revised Edition, Routledge, London and New York, 1994) at 43ff.
made particularly clear by the issues faced by multilingual jurisdictions, where translation is not only implicit but an explicit part of the legislative process.
The concept of multilingualism is in itself a complex one, but for our purposes it is sufficient to define it as “the use of two or more languages by an individual speaker or a community of speakers.”4 A multilingual legislature is one which ‘speaks’ two or more languages in the drafting and interpretation of its legislation, as opposed to a monolingual legislature which employs only one language. The professed aim of most bilingual and multilingual legislatures is to provide better access to justice for their citizens, particularly those who are part of a minority or colonised culture, although the practical outcome tends more to reflect an underlying aspiration towards community amalgamation than it does a desire to give everyone a better grasp of the law.5
In most cases, multilingualism is achieved by means of an official language policy, which determines what languages will be part of the legislative repertoire and establishes the ground rules of their relationship to one another. The precise nature and content of this language policy is, of course, context-dependent, and runs the gamut from jurisdictions such as the European Union, which produces legislation in most or all of the languages of its various member states, to those like Canada and Hong Kong, which each employ only two
4 Janny Leung “Statutory Interpretation in Multilingual Jurisdictions: Typology and Trends.” (2012) 33(5) Journal of Multilingual and Multicultural Development 481 at 482.
5 Leung, “Statutory Interpretation”, at 481; By contrast, Ruth Sullivan argues that multilingualism assumes greater linguistic unity within the general population, and may thus be seen as promoting community more than accessibility (“The Challenges of Interpreting Multilingual, Multijural Legislation.” (2003) 29 Brook J Int’l L 985 at 1009.
languages. To be truly multilingual, however, mere explanatory translation is not sufficient; a legislature must give equal weight to the legal text of each language.6 In Canada, for example, s 133 of the Constitution Act 1867 establishes that members of the Dominion of Canada have the right to use English or French in the Parliament of Canada and the Legislature of Quebec, as well as in any proceedings before a federal court or a court of Quebec (s 133(1)).7 It also provides that the Acts of the Canadian Parliament and the Legislature of Quebec shall be printed and published in both languages (s 133(2)). This was later reinforced by the Official Languages Act 1969 and again in 1988, the latter of which provides not only that “all Acts of Parliament shall be enacted, printed and published in both official languages” (s 6) but that both versions are equally authentic and have equal legal force (s 13).8
By contrast, in Hong Kong prior to 1989 all legislation was enacted in English. After the handover to China, the Hong Kong Official Languages Ordinance amended this to establish both English and Chinese as official languages in Hong Kong (s 3(1)), and to require that all legislation be enacted in both English and Chinese (s 4(1)). Since then, Hong Kong statute law has become fully bilingual, and under s 10B of the Hong Kong Interpretation and General Clauses Ordinance both the English and Chinese texts have (in theory) the same authenticity and legal force.9 Despite this, however, because most of the Chinese texts were post-hoc translations of existing laws in English,
6 Leung, “Statutory Interpretation”, at 482–483.
7 Originally the British North America Act 1867, renamed in 1982. University of Ottawa, “The Constitution Act of 1867 and the Language Question” Site for Language Management in Canada <http://www.slmc.uottawa.ca/> .
8 For discussion as to how this differs in each province, see Sullivan, “Challenges” at 1006.
9 Deborah Cao “Chapter 3: Judicial Interpretation of Bilingual and Multilingual Laws: A European and Hong Kong Comparison” in Dr. J. Jemielniak and Dr.
P. Mikłaszewicz (eds), Interpretation of Law in the Global World: From Particularism to a Universal Approach (Springer, Heidelberg, 2010) at 80.
where the two versions differ the courts will frequently regard the Chinese text as being in error, and defer instead to the English meaning:10
if the Ordinance was initially enacted in English, the English text was the original official text from which the Chinese text was subsequently prepared and declared authentic. In ascertaining the ordinance’s legal meaning, the English text should be taken as more accurately reflecting the legislature’s intent at the time it was originally enacted.
The meaning of the English text is thus prioritised over the Chinese one, suggesting that, while a legislature may be multilingual by virtue of the fact that it employs multiple languages, some legislatures are (for lack of a better phrase) more multilingual than others.
For any degree of legal multilingualism, however, translation plays an explicit role in the creation of legislation. There are two basic methods by which multilingual legislation may be drafted: first, through the drafting of a statute in one or more of the official languages, which is translated into the other(s) after it has reached its final form, and second, through co-drafting.11 Co-drafting is a difficult and involved process, and relatively uncommon in practice. In Canada, public law statutes are typically co-drafted in English and French simultaneously, a unique system which was adopted by the Department of Justice in 1978. Since then, all government bills have been drafted by teams of two drafters, typically including one English-speaker (trained in
10 HKSAR v Lau San Ching and Others HCMA 98/2002; See also Chan Fung Lan v Lai Wai Chuen HCMP4210/1996.
11 Michael J B Wood “Drafting Bilingual Legislation in Canada: Examples of Beneficial Cross-Pollination Between the Two Language Versions.” (1996) 17(1) Stat LR 66 at 69.
common law) and one French-speaker (trained in civil law).12 In this case, translation is close to immediate and serves as a practical tool to ensure that both the English and the French texts resemble one another as closely as possible. Both drafters share the same information, and work in close collaboration together, such that each version of the text is to some degree informed by the other.13 It is therefore necessary for both drafters to be sufficiently fluent in both languages to be able to participate in meaningful discussion as to the implications and interpretations of both legal texts. They are also assisted by bilingual support departments, whose job is to approve both versions of the legislation.14 While in practice one or the other of the drafters may take the lead for pragmatic reasons, and finish their draft first before passing it on to the other for subsequent translation and mutual revision, both versions are to some extent created together on computer screens in the co-drafting rooms.15 The end result is a bilingual text which is enacted as a coherent whole, and the subsequent promulgation of the legislation in both languages at the same time.
More commonly, other jurisdictions tend to use translation as a means of replicating an already-drafted text into one or more other official languages, similar to the procedure in Hong Kong. In the European Union, for example, legislation is first proposed by the European Commission, and a draft prepared by technical experts in English and/or French, depending on the language of the department in which the law is made.16 It is then submitted to other Commission departments for internal consultation, where the Commission’s Legal Service and legal reviewers examine it for appropriate form and presentation. Only then is it translated into the rest of the official
12 Lionel A Levert “Bilingual and Bijural Legislative Drafting: To Be or Not to Be?” (2004) 25(2) Stat LR 151 at 155.
14 Ibid at 156.
15 Levert, “Bilingual and Bijural Legislative Drafting”, at 155–156.
16 Cao, “Judicial Interpretation”, at 74.
languages by the Directorate-General for Translation, and, after further revision, submitted to the European Parliament and Council for internal pre-adoption procedures.17 The effectiveness of the resulting legislation therefore depends predominantly on the ability of the European Court of Justice to “harmonise the different versions of its multilingual texts.”18
Explicit in the methodology of multilingual drafting, then, is a clear process of negotiation for meaning. Translation is seldom a one-to-one equivalence between one language and the next: not only the formal structure but also the conceptual content of words and phrases requires frequent adjustment to accommodate linguistic differences.19 Michael Wood, in his survey of ‘beneficial cross-pollination' between French and English drafting in Canada, gives the example of the English word “any” as a source of potential ambiguity, as it can mean either “all” or “one” depending on how it is used in a sentence.20 When translated into French, the English rendering:
(1) The report shall include any document specified in the schedule.
could become either of the following alternatives: 21
(1) La rapport comprend l'un des document énoncés à l’annexe (“one of the documents”); or
(2) La rapport comprend les documents énoncés à l’annexe (“the documents”).
18 Ibid at 75.
19 Wood, “Drafting Bilingual Legislation”, at 70.
21 Wood, “Drafting Bilingual Legislation”, at 70.
French in this sense is more precise than English, using articles to indicate number in a way that renders the two potential meanings of “any” disparate rather than convergent. Similarly, the rules of grammar and syntax in French may require that which is elided in an English statute to be made explicit in the French version. Thus:
(1) The rules may provide for the refusal to grant or renew or the suspension or revocation of a permit.
(1) Les regies peuvent prevoir le refus de délivrer ou de renouveler un permis ou la suspension ou la révocation d'un permis.
What this ultimately points to are the underlying conceptual and cultural differences between the languages themselves. Semiotic theory suggests that the conception of language as constituting a one-to-one equivalence between signifier (that which is said) and signified (that which is meant) fails to account for a key component of linguistic meaning which is contingent on factors extrinsic to the sign itself. Roland Barthes’ conception of three-dimensional semiotics posits that signs are made up not only of signifier and signified but also a third dimension, which he calls “myths.”23 These myths serve as a second- order semiological system or metalanguage (language about language) insofar as the sign consisting of the straight-forward linguistic term and its associated concept serves as a mere signifier for some additional cultural implication making up the myth itself.24
22 Ibid at 72.
23 Roland Barthes Mythologies (Vintage Books, London, 2009) at 131–187.
Barthes gives the example of a phrase in a Latin textbook: quia ego nominor leo.25 The literal translation is simply “because my name is lion,” taken from a text by Phaedrus. However, it has in context an additional meaning as a grammatical example, intended to demonstrate a specific grammatical rule (the agreement of the predicate).26 Meaning, for Barthes, is thus inherently dependent on cultural and historical context, which the myth acts to naturalise or de-politicise, presenting as fact what is in actuality a system of value.27 A similar principle inheres in the syntactic and grammatical structures of language, insofar as the way in which a language system orders itself and prioritises content (subject- object-verb versus subject-verb-object, topic-oriented sentences versus subject-oriented sentences and so forth) necessarily says something about the priorities and perspectives of the culture to which it belongs.28
Taking this to its logical conclusion, it may be said that language shapes thought and thought language, to the extent that some linguists maintain it is impossible to acquire a second language with sufficient facility to think the same way as a native speaker, because the backgrounds and processes are simply too different and too subtle to be learned.29 The act of transferring even the most straightforward information from, for example, English to French, or vice versa, thus inevitably becomes deeply complex. In some cases, the differences between two languages may be irreconcilable,30 suggesting that
25 Ibid at 138–140.
26 Ibid at 139.
27 Ibid at 168–172.
28 Lourdes Ortega Understanding Second Language Acquisition (Routledge, London and New York, 2013) at 44–46.
29 Ibid at 47.
30 Chinese, for example, has no articles and a different system of plural markers, causing problems with interlingual translation: Deborah Cao “Interlingual Uncertainty in Bilingual and Multilingual law.” (2007) 39 Journal of Pragmatics 69 at 79.
translation must incorporate an element of interpretation or ‘transformation’ beyond mere formal or semantic equivalence.
Of course, this disconnect between appearance and sense can be, and frequently is, exploited in a multilingual context. Deborah Cao, for example, makes note of the ambiguous use of the word ‘sorry’ in the United States response to China in April, 2001.31 After a US spy plane collided with a Chinese jet and was forced to land in Lingshui, the Chinese Government demanded an apology, feeling that the incident reflected badly on its sovereignty and would affect its standing in the eyes of its people. The American Government, meanwhile, insisted it would not apologise as it had done nothing wrong.32 In the end, the US wrote to China expressing ‘sincere regret’ over the missing pilot and aircraft, adding that it was ‘very sorry’ that the US plane had entered China’s airspace and landed in China without the verbal clearance of the authorities. As Cao herself writes, “‘sorry’ in English can mean both ‘regret’ and ‘apology.’ In translating ‘sorry’ into Chinese, two different words have to be used, vihan for the former and daoqian for the latter. It seems that the American ‘sorry’ letter intended the former meaning ‘regret,’ but the Chinese government chose to interpret and translate the second meaning ‘apology.’”33
Likewise, at the level of drafting, there may be good reasons to allow certain elements of vagueness or generality to remain in the legal text.34 For the law to be applicable to specific facts, however, it requires that some consensus as to meaning be reached, and the influence of translation can thus also be seen in the methods of multilingual
31 Cao, “Interlingual Uncertainty”, at 71.
33 Cao, “Interlingual Uncertainty”, at 71.
34 See LL Fuller The Morality of Law (Revised Edition, Yale University Press, New Haven, 1969) at 33–94.
interpretation. In Classical antiquity, early translation theory suggested that poetic translation should be approached by the translator non verbum de verbo, sed sensum exprimere de sensu,35 which included adapting the original text into a unique piece of literature in its own right. Yet embracing free interpretation in this way became problematic when the text to be translated was something of great inherent authority, such as Biblical scripture. In his translation of the Vulgate, St. Jerome suggested that Biblical language possessed “a numinous character far more important than its communicative function”36 and therefore ought to be preserved as closely as possible. As translation of scriptural texts into the vernacular became more common, a balance had to be struck between the evangelical necessity of translation on the one hand, and the preservation of the inherent sacredness of the text on the other: the “word-for-word” or “sense-for-sense” dichotomy which continues to be debated amongst translators today.37
If law is indeed “something close to secular scripture”,38 then it is hardly surprising to find that this tension also resides in the ‘translation’ or interpretation of legislation, both in monolingual and in multilingual contexts. As noted earlier, the Equal Authenticity Rule by which most multilingual jurisdictions abide entails that each language version has equal status as law. In addition, most multilingual jurisdictions follow some variation of the Shared Meaning Rule, the presumption that all versions of a multilingual legal text share the same meaning unless otherwise demonstrated.39 Taken together, these two interpretative rules essentially entail that “an interpreter cannot know the substance of the law declared by Parliament until he or she has considered [all]
35 That is, “expressing not word for word, but sense for sense”: Susan Bassnett Translation Studies (Revised Edition, Routledge, London and New York, 1994) at 44.
36 Stanton, “The Culture of Translation”, at 112.
37 Bassnett, “Translation Studies”, at 45–46.
38 Roderick A MacDonald “Legal Bilingualism” (1996) 42 McGill LJ 119 at 132.
39 Sullivan, “Challenges”, at 1012ff.
versions and resolved any discrepancies between them.”40 The literal interpretation of a word or phrase may thus be at odds with its ‘sense- for-sense’ interpretation, entailing that one must read across the various versions of the text to divine the ultimate (perhaps even transcendental) meaning of which the ‘true’ legal text is comprised.41
In Fonden Marselisborg Lystbådehavn v Skatteministeriet, for example, the question arose in relation to Article 13(b)(2) of the Six Directive as to what was encompassed by the word ‘vehicle.’42 A Danish company, which provided leases to boats, argued that a boat was not a ‘vehicle’ and therefore did not fall within the provision. While some versions of the legislation supported this distinction, however, using a term which related specifically to land-based transport, a substantial number used generalised terms which encompassed all forms of transportation, including boats. In its decision, the Court held that the word ‘vehicle’ as it ought to be understood took the second meaning, citing what Cao refers to as “the purpose and scheme of the relevant law.”43 In a similar vein, the European Union has also made it clear that words can have their own independent meanings in Community law incorporating all the extant language versions, and in cases of interlingual ambiguity the established solution is to favour that interpretation which ensures the effectiveness of the legislature's purpose—that is, one which follows the spirit rather than the letter of the law.44
The chief insight of bilingual and multilingual legislation therefore lies in its ability to demonstrate that linguistic uncertainty in the law is not
40 Ibid at 1007.
41 Bassnett, “Translation Studies”, at 44–45.
42 Case C-428/02 Fonden Marselisborg Lystbådehavn v Skatteministeriet  ECR I-1527.
43 Cao, “Judicial Interpretation”, at 75.
44 Ibid at 77.
confined solely to any one language. While law does depend on precision and unitary meaning to operate, language is intrinsically indeterminate and requires the use of rules of construction to resolve its inherent ambiguities, the more so when it involves establishing meaning between two or more different languages.45 Multilingual legislation may involve a greater degree of complication than monolingual legislation, but the two are in many ways deeply connected, inasmuch as the latter does explicitly what the former does implicitly: it engages in an explicit process of translation or dialogue between two separate systems of communication, both in the process of drafting and in the process of interpretation. In this way, a deeper understanding of multilingual jurisdictions reveals that the law itself is a constant process of negotiation, from the moment of its inception through its promulgation and application by the courts. Whether monolingual or multilingual, therefore, the law is in some sense a process of translation.
45 Ibid at 84.